Cloud Stack allows you to build and publish cloud-based structures from templates, which are used as part of SEO and link-building workflows. Through MCP, your AI assistant can create stacks, publish them, and track their status.

This article explains how to create, build, and manage cloud stacks using MCP.

## 🧠 What You Can Do

Using MCP with Cloud Stack, you can:

- Create cloud stacks from templates
- Build cloud stack structures
- Publish cloud stacks
- Track build and publication status
- Monitor quota usage

👉 Cloud stacks are used as part of broader SEO workflows, particularly in link-building strategies.

## ✅ Prerequisites & Plan Requirements

Before using Cloud Stack via MCP, please note:

- MCP access to Cloud Stack tools is included in plans that have the Cloud Stack add-on or feature enabled. If you are unsure whether your plan includes Cloud Stack MCP access, contact your account manager or Search Atlas support.
- Certain Cloud Stack capabilities (such as specific providers or higher-volume publishing) may be restricted by plan tier or require special enablement on your workspace.
- If an MCP endpoint is not enabled for your plan, your AI assistant will return an access or permission error when invoking the tool.

## ☁️ Supported Cloud Providers & Templates

Cloud stacks are created from provider templates. Search Atlas currently supports 15+ cloud providers, including recently added git-based and sandbox providers such as:

- GitLab Pages
- AWS Amplify
- SourceForge
- Glitch
- CodeSandbox

👉 When you ask your AI assistant to create a cloud stack, you can specify which provider template to use. If you do not specify one, the assistant may list available providers or pick a default based on your workflow.

## ⚙️ Step-by-Step: Create a Cloud Stack

## Step 1

Open your MCP-connected AI assistant

## Step 2

Ask to create a cloud stack

Examples:

- “Create a cloud stack”
- “Set up a cloud stack from a template”
- “Create a cloud stack on GitLab Pages” (or any supported provider)

## Step 3

The assistant creates the stack from the chosen provider template

👉 Stacks are always created from one of the supported provider templates listed above.

## Step 4

Confirm stack details

⚠️ **Note (upcoming change):** The separate `cs_create` and `cs_build` MCP tools may be merged into a single `cs_create` action. Once that change ships, the Create and Build steps below may be combined into one command — your AI assistant will handle creation and build in a single step.

## 🏗️ Step-by-Step: Build the Cloud Stack

💡 **Build vs. Publish:** Building compiles and prepares the stack structure (assembling files, configuration, and template content) so it is ready to deploy. Publishing takes the built stack and makes it live/publicly accessible at its provider URL. Both steps are currently required; they may be merged into a single step pending an upcoming release.

## Step 1

Ask your assistant:

- “Build my cloud stack”

## Step 2

The assistant initiates the build process

## Step 3

Wait for completion

👉 Building prepares the stack for publishing — the stack is not yet live at this point.

## 📤 Step-by-Step: Publish the Cloud Stack

⚠️ **Important — publishing is potentially immediate and not easily reversible.** Once a stack is published, it becomes live at its provider URL. Always review your stack content and configuration before issuing a publish command.

## Step 1

Ask your assistant:

- “Publish my cloud stack”

## Step 2

The assistant prepares the publication

## Step 3

Approve the action (if required)

When approval is triggered, your AI assistant will display a confirmation dialog or return an MCP response that requires acknowledgment before the publish call is executed. To proceed, reply with an explicit confirmation such as “Yes, publish” or click the approve button in your assistant's confirmation prompt. To cancel, reply “No” or “Cancel” — the stack will remain unpublished.

Whether the approval prompt appears depends on your MCP client's configuration: some clients are configured to auto-approve tool calls, in which case the publish may execute immediately. If you want to ensure a manual approval step every time, configure your MCP client to require confirmation for Cloud Stack publish actions.

## Step 4

The stack is published

👉 Publishing makes the cloud stack active and accessible at its provider URL.

## 📊 Step-by-Step: Track Status

## Step 1

Ask your assistant:

- “Show my cloud stack status”

## Step 2

The assistant retrieves status data

## Step 3

Review:

- build status — whether the stack has been compiled and prepared for publishing (e.g., pending, in progress, completed, failed)
- publication status — whether the stack is live at its provider URL (e.g., unpublished, publishing, published, error)
- progress updates — incremental updates while build or publish operations are running

## Step 4

Interpret the result and take next steps:

- If build status shows failed, ask your assistant for the error details and re-run the build after addressing the issue.
- If publication status shows published, your stack is live — verify it at the provider URL.
- If status appears stuck, wait a moment and ask your assistant to refresh the status.

## 📈 Monitoring Quota Usage

Cloud Stack actions consume Search Atlas quota. To check your usage via MCP:

- Ask your assistant: “Show my Cloud Stack quota” or “How much Cloud Stack quota do I have left?”
- The assistant will return your current quota balance and usage for Cloud Stack operations.

Quota limits depend on your plan and any add-ons. Creating, building, and publishing stacks each may consume quota, and certain providers may consume more than others.

⚠️ **Note:** The quota displayed in the Search Atlas dashboard may not always reflect what is accessible via MCP. If a specific endpoint is restricted on your plan, the MCP tool may return an access error even when the dashboard shows available quota. Contact support if you see a mismatch.

## ⚠️ Important Notes

1. Creating and building stacks may consume Search Atlas quota
2. Publishing may involve additional resource usage
3. Actions may require approval depending on execution
4. Cloud stacks are part of broader SEO workflows
5. Publishing via MCP is potentially immediate and not easily reversible — always review your stack before issuing a publish command

## 🔁 How This Fits Into MCP Workflows

A typical workflow:

1. Create cloud stack
2. Build stack
3. Publish stack
4. Track status

👉 This supports infrastructure-based SEO execution.

## 🧠 Best Practices

1. Confirm configuration before building
2. Review stack content before issuing a publish command — publishing is potentially immediate
3. Monitor status after publishing
4. Use cloud stacks as part of larger SEO strategies
5. Track results alongside other SEO tools

MCP for Cloud Stack allows you to create and deploy cloud-based SEO structures through conversation. By building and publishing stacks from templates, your AI assistant can execute infrastructure-level SEO actions using the same capabilities available in Search Atlas.